2026-03-05 ヒューストン大学（UH）
Handheld radar used to detect hidden damage in buildings
＜関連情報＞
- https://www.uh.edu/news-events/stories/2026/march/03052026-hoskere-cold-steel-radar-defects.php
- https://ascelibrary.org/doi/abs/10.1061/JCCEE5.CPENG-7168
地中レーダーとビジョン基礎モデルを統合した隠蔽冷間成形鋼構造部材と損傷評価 Concealed Cold-Formed Steel Structural Members and Damage Assessment Integrating Ground Penetrating Radar with Vision Foundation Model
Muhammad Taseer Ali, S.M.ASCE, and Vedhus Hoskere, Ph.D., M.ASCE
Journal of Computing in Civil Engineering Published:Jan 31, 2026
DOI:https://doi.org/10.1061/JCCEE5.CPENG-7168
Abstract
Timely and accurate assessment of concealed cold-formed steel (CFS) structural members is essential for ensuring the integrity and longevity of buildings. Traditional inspection methods require partial or complete removal of cladding, making the process labor-intensive, costly, and inefficient. To address these limitations, we introduce a novel framework that integrates ground penetrating radar (GPR) with a state-of-the-art large-scale vision foundation model, InternImage, for automated detection of CFS members and damage. Our work presents three key contributions: (1) introducing the use of GPR for nondestructive condition assessment of concealed CFS members; (2) developing and curating CFS-GPR, a data set containing diverse member orientations, damage types, and cladding combinations to evaluate model performance and optimal annotation techniques; and (3) introducing GPR-CutMix, a novel augmentation method that enhances model generalizability to unseen data simulating realistic variations in member spacing. Experiments are first carried out on data collected from a custom laboratory setup for model training, hyperparameter tuning, and GPR-CutMix validation. Finally, we demonstrate the impact of GPR-CutMix on the model’s ability to generalize across data from real buildings with different cladding configurations when trained on laboratory data only. These findings highlight the potential of our framework to advance the CFS structural inspection methods by providing a rapid, reliable, and scalable approach for damage detection, ultimately improving building maintenance and rehabilitation.
